Dielectric resonators are used in many circuits, particularly microwave circuits, for concentrating electric fields. They can be used to form filters, oscillators, triplexers and other circuits. The higher the dielectric constant of the dielectric material out of which the resonator is formed, the smaller the space within which the electric fields are concentrated. Suitable dielectric materials for fabricating dielectric resonators are available today with dielectric constants ranging from approximately 10 to approximately 150 (relative to air). These dielectric materials generally have a magnetic constant of 1, i.e., they are transparent to magnetic fields.
FIG. 1 is a perspective view of a typical dielectric resonator of the prior art. As can be seen, the resonator 10 is formed as a cylinder 12 of dielectric material with a circular, longitudinal through hole 14. Individual resonators are commonly called “pucks” in the relevant trades. While dielectric resonators have many uses, their primary use is in connection with microwaves and, particularly, in microwave communication systems and networks.
As is well known in the art, dielectric resonators and resonator filters have multiple modes of electrical fields and magnetic fields concentrated at different center frequencies. A mode is a field configuration corresponding to a resonant frequency of the system as determined by Maxwell's equations. In a dielectric resonator, the fundamental resonant mode frequency, i.e., the lowest frequency, is the transverse electric field mode, TE01 (or TE, hereafter). Typically, it is the fundamental TE mode that is the desired mode of the circuit or system into which the resonator is incorporated. The second mode is commonly termed the hybrid mode, H11 (or H11, hereafter). The Ha, mode is excited from the dielectric resonator, but a considerable amount of electric field lays outside the resonator and, therefore, is strongly affected by the cavity. The H11 mode is the result of an interaction of the dielectric resonator and the cavity within which it is positioned. The H11 mode field is orthogonal to the TE mode field. There also are additional higher modes. Typically, all of the modes other than the mode of interest, e.g., the TE mode, are undesired and constitute interference. The H11 mode, however, typically is the only interference mode of significant concern. The remaining modes usually have substantial frequency separation from the TE mode and thus do not cause significant interference with operation of the system. The H11 mode, however, tends to be rather close in frequency to the TE mode. In addition, as the frequency of the TE mode is tuned, the center frequency of the TE mode and the H11 mode move in opposite directions to each other. Thus, as the TE mode is tuned to increase its center frequency, the center frequency of the H11 mode inherently moves downward and, thus, closer to the TE mode center frequency. By contrast, the third mode, commonly called the H12 mode, not only is sufficiently spaced in frequency from the TE mode so as not to cause significant problems, but, in addition, it moves in the same direction as the TE mode responsive to tuning.
FIG. 2 is a perspective view of a cross-coupled dielectric resonator filter 20 of the prior art employing a plurality of dielectric resonators 10. The resonators 10 are arranged in the cavity 22 of a conductive enclosure 24. The conductive enclosure 24 typically is rectangular, as shown in FIG. 2. Microwave energy is introduced into the cavity via an input coupler 28 coupled to a cable, such as a coaxial cable. The energy may then be coupled to a first resonator (such as resonator 10a) using a coupling loop. Conductive separating walls 32 separate the resonators from each other and block (partially or wholly) coupling between physically adjacent resonators 10. Particularly, irises 30 in walls 32 control the coupling between adjacent resonators 10. Conductive walls without irises generally prevent any coupling between the resonators separated by the walls, while walls with irises allow some coupling between these resonators. Specifically, conductive material within the electric field of a resonator essentially absorbs the field coincident with the material and turns it into a current in the conductor so that the field does not pass through to the other side of the wall. In other words, conductive materials within the electric fields cause losses in the circuit.
Conductive adjusting screws (not shown) coupled to the enclosure may be placed in the irises to further affect the coupling of the fields between adjacent resonators and provide adjustability of the coupling between the resonators, but are not used in the example of FIG. 2. When positioned within an iris, a conductive screw partially blocks the coupling between adjacent resonators permitted by the iris between them. Inserting more of the conductive screw into the iris reduces coupling between the resonators while withdrawing the conductive screw from the iris increases coupling between the resonators.
A cross-coupler 34 having a metal probe 34a extending through a non-conductive bushing 34b is used to couple resonators separated by walls without irises to obtain more optimum filter transfer functions. The non-conductive bushing 34b electrically isolates the probe 34a from the enclosure 24 so that electric fields coincident to the probe 34a are not absorbed by the walls of the enclosure, but rather are passed from one end of the probe 34a to the other for coupling resonators adjacent the ends of the probe 34a. 
By way of example, the field of resonator 10a couples to the field of resonator 10b through iris 30a, the field of resonator 10b further couples to the field of resonator 10c through iris 30b, and the field of resonator 10c further couples to the field of resonator 10d through iris 30c. In addition, the field of resonator 10a further couples to the field of resonator 10c through cross-coupler 34. Wall 32a, which does not have an iris or a cross-coupler, prevents the field of resonator 10a from coupling with the physically adjacent resonator 10d on the other side of the wall 32a. 
One or more metal plates 42 may be positioned adjacent each resonator to affect the field of the resonator to set the center frequency of the filter. Particularly, plate 42 may be mounted on a screw 44 passing through a top surface (not shown) of the enclosure 24. The screw 44 may be rotated to vary the spacing between the plate 42 and the resonator 10 to adjust the center frequency of the resonator. A coupling loop connected to an output coupler 38 is positioned adjacent the last resonator 10d to couple the microwave energy out of the filter 20. Signals also may be coupled into and out of a dielectric resonator circuit by other methods, such as microstrips positioned on the bottom surface 44 of the enclosure 24 adjacent the resonators. The sizes of the resonators 10, their relative spacing, the number of resonators, the size of the cavity 22, the size of the irises 30, and the size and position of the metal plates 42 all need to be precisely controlled to set the desired center frequency of the filter, the bandwidth of the filter, and the rejection in the stop band of the filter. More specifically, the bandwidth of the filter is controlled primarily by the amount of coupling of the electric and magnetic fields between the resonators. Generally, the closer the resonators are to each other, the more coupling between them and the wider the bandwidth of the filter. On the other hand, the center frequency of the filter is controlled in large part by the size of the resonator and the size and the spacing of the metal plates 42 from the corresponding resonators 10.
In an alternative prior art cross-coupled dielectric resonator filter, coaxial cables are used to couple resonators that are separated by walls without irises and/or are not adjacent to one another. A detailed discussion of cross-coupled dielectric resonators is found in U.S. Pat. No. 5,748,058 to Scott entitled CROSS COUPLED BANDPASS FILTER, incorporated fully herein by reference.
Prior art cross-coupled dielectric resonator filters have limited frequency bandwidth performance. The maximum frequencies at which they can perform effectively are typically limited to about 55 to 60 GHz. The effective bandwidth range of prior art cross-coupled dielectric resonator filters is typically on the order of 3 to 20 MHz. In particular, the bandwidth is restricted because the coupling between resonators is limited.
Prior art resonators and the cross-coupled resonator circuits made from them have many drawbacks. For instance, as a result of the positions of the fields of the resonators, prior art resonators have limited ability to couple with other resonators (or with other microwave devices such as loop couplers and microstrips). That is why filters made from prior art resonators have limited bandwidth range. Further, prior art cross-coupled dielectric resonator circuits rely on probes or coaxial cables for cross-coupling, and filter poles may have to be laid in a zig-zag manner, which put significant constraints on filter performance. In addition, prior art cross-coupled dielectric resonator circuits such as the filter 20 shown in FIG. 2 suffer from poor quality factor, Q, due to the presence of separating walls and coupling screws between adjacent resonators. Q essentially is an efficiency rating of the system and, more particularly, is the ratio of stored energy to lost energy in the system. The fields generated by the resonators pass through all of the conductive components of the system, such as the enclosure 24, plates 42, and internal walls 32 and inherently generate currents in those conductive elements. Those currents essentially comprise energy that is lost to the system.
Furthermore, the volume and configuration of the conductive enclosure 24, substantially affects the operation of the system. The enclosure minimizes radiative loss. However, it also has a substantial effect on the center frequency of the TE mode. Accordingly, not only must the enclosure be constructed of a conductive material, but it must be very precisely machined to achieve precise coupling and the desired center frequency performance, thus adding complexity and expense to the fabrication of the system. Even with very precise machining, the design can easily be marginal and fail specification.
Furthermore, prior art resonators have poor mode separation between the desired TE mode and the undesired H11 mode.
FIGS. 3A and 3B illustrate typical locations for the TE and H11 modes, respectively, in a typical prior art cylindrical resonator 10. As shown, the electric field lines of the TE mode field 50 are circular, oriented parallel to the circular dimension of the resonator cylinder 12 and concentrated around the circumference of the resonator 10, with some of the field within the resonator and some of the field without the resonator. A portion of the field is outside the resonator for purposes of coupling between the resonator and other microwave devices (e.g., other resonators or input/output couplers).
The H11 mode field 60 is orthogonal to the TE mode. The electric field lines of field 60 form circles parallel to the page in FIG. 3B and are concentrated near the surface of the cylinder 12. It is very difficult to physically separate the H11 mode from the TE mode. Accordingly, methods for suppressing the H11 mode have been developed in the prior art. For instance, metal strips 70 such as illustrated in FIG. 4 have been placed on the surface of the resonators to suppress the H11 mode by causing its tangential electric field to be zero at the metal strips 70, effectively causing the suppression of the mode because its maximum field strength is located near the metal strips. In practice, while this technique for suppressing the H11 mode is relatively effective in terms of suppressing the H11 mode, it also typically suppresses the TE mode significantly. In theory, the effect on the TE mode should be insignificant, but experiments show that this is not the case in the real world and that this method for H11 suppression actually significantly affects Q for the TE mode. Experiments show that this technique typically might cause losses of about half of the power of the TE mode, thus substantially reducing the Q of the resonator and the overall system in which it is employed.
Accordingly, it is an object of the present invention to provide improved cross-coupled dielectric resonator circuits.
It is another object of the present invention to provide improved cross-coupled dielectric resonator filters.
It is another object of the present invention to provide cross-coupled dielectric resonator circuits in which the H11 mode is substantially suppressed or eliminated.
It is yet a further object of the present invention to provide cross-coupled dielectric resonator circuits that are easily tunable.
It is one more object of the present invention to provide cross-coupled dielectric resonator circuits with more effective coupling and cross-coupling than in the state of the art.
It is a further object of the present invention to provide cross-coupled dielectric resonator circuits with improved Q factors.
It is yet a further object of the present invention to provide cross-coupled dielectric resonator circuits with improved layouts.
It is one more object of the present invention to provide cross-coupled dielectric resonator filters having compact packaging.